Learning Analytics and Outcomes Template

Completion rate, broken down by the lesson where learners drop off. Completion rate is actionable: if Module 3 Lesson 2 has a 40% drop-off, you know exactly where to investigate. Combine it with assessment scores to distinguish between "learners who finish and learn" vs "learners who click through without learning." Track both, but start with completion funnels.

Pre/post assessment. Give learners an assessment before the course (diagnostic) and the same or equivalent assessment after. The score difference is your learning gain. This is the most rigorous measurement that does not require external data. For deeper outcome measurement, survey learners and their managers 90 days after completion to assess behavior change. Kirkpatrick Level 3 (behavior) is the gold standard for corporate training ROI.

For self-paced online courses: 15-30% is typical across the industry (MOOCs average 5-15%). For cohort-based courses with deadlines: 60-80%. For employer-mandated training: 85-95%. If your self-paced completion rate is above 30%, you are outperforming most platforms. Focus optimization efforts on the first 3 lessons (where most drop-off occurs) and the assessment gates (where anxiety causes abandonment).

For engagement metrics (completion rates, time-on-task, session frequency): useful from day one with 50+ learners. For assessment item analysis (difficulty index, discrimination): need 100+ responses per item for stable estimates. For predictive models (at-risk identification, outcome prediction): need 500-1000+ learners with full funnel data. Start simple with descriptive dashboards and add predictive features as your data grows.

Build a custom event system for v1 and add xAPI compliance later. xAPI (Experience API) is a standard for learning event data, but its specification is complex and most internal analytics needs are served faster by a simple custom event taxonomy. If you need to interoperate with external LMS platforms or comply with enterprise requirements, add an xAPI export layer on top of your custom events. Do not let xAPI complexity slow down your initial analytics implementation.